From: Lee Corbin (lcorbin@tsoft.com)
Date: Thu Mar 06 2003 - 00:38:57 MST
Hal writes
> > It remains for us only to determine how much run time people
> > ---including ourselves---are to get. And of course, any
> > answer failing to specify the maximum possible fails to
> > attain the highest morality.
>
> Yes, perhaps you can re-style yourself The Speaker for the
> Underinstantiated. Rather than rescuing people from non-existence,
> you can strive to maximize the total quantity of sentient existence.
Nope. It's a no-go. Everyone is underinstantiated, and
it makes no sense to be "speaker for everyone". Before,
I was reacting to the arrogance of the instantiated, who
so often dismissed away the needs of the uninstantiated,
who in turn themselves seemed to have no advocate.
> One of the paradoxes that I have always struggled with is this: if you
> run exactly the same conscious program twice, does it matter? Does it
> increase the "measure" or "probability" of that conscious experience?
> Do I do good by re-running someone's pleasant experience, and harm by
> re-running a bad one?
Yes, yes, and yes. At moments like this, I see with utter
clarity the benefits of repeated experience. I feel only
sorry that others have some kind of unconscious goals or
aspects of their self-images that make repeated experience
seem repugnant. But it is late, and I'm being uncharitable.
> The reason it's a paradox is because re-runs are particularly prone to
> philosophical slippery-slope arguments where we blur the lines between an
> "actual" re-run and what amounts to a picture of one. I won't recap the
> arguments here, they have been discussed ad infinitum in the archives.
It's utterly crazy to think that a picture of a suffering child suffers.
And how brilliant and professional philosophers like Hilary Putnam can
*find* every possible child or civilization in every stone defies belief.
At moments like these, endorsing a Theory of Dust---in which information
never flows, but is only statically represented---seems utterly wrong.
> The point is that only re-runs are susceptible to these arguments;
> fresh runs doing novel calculations cannot be replaced by pictures of
> themselves because the pictures won't exist until the calculation is done.
Yes! Now THAT is exactly right. But...
> This has led me to consider the possibility that re-runs don't count,
> that they carry no experiential (and therefore moral) weight.
How do you go suddenly to that? A calculation is a real-time
*****process*****, and as such it implements information flow.
Whether a picture exists, before a calculation, during a calculation,
or after a calculation is completely unimportant. I tell you,
Mao Ze Dong got satisfaction only from *seeing* all those pictures
of himself, never even once from being any of them.
> However, the notion of many-worlds at Tegmark's level 1 and 2 suggests
> that we do have to consider identical universes to add to the measure
> and probability of the experiences they contain. We have all these
> identical Earths, and for probability to work right we need to assume
> that the total probability for a given experience is proportional to
> the fraction of Earths where that experience happens, even though many
> of them will be identical. Therefore identical calculations must be
> counted and added separately in considering their contribution to the
> overall probabilities of experience.
That is the horn of your "paradox" that I totally agree with.
Consider even the argument from philosophical continuity:
We have two almost identical Hal Finneys in separate solar
systems, and (somehow it's obvious to me that) each is getting
great benefit from his life. We change one atom. Now suddenly
they're identical, and the total benefit goes from two to one?
Preposterous.
But you are right. This has been hashed through many times
in the archives. Nonetheless (at this late hour) I express
the hope that everyone will come to see that I have been
right. ;-) The level one infinite universe practically
proves it. Since there are other Lees exactly like me
only 10^10^29 meters away, the sudden demise of any particular
one shouldn't be of any concern if repeated experience had no
value. But we know better.
Now we have yet another case where the pointer this* is
multiple valued, and we cannot blithely speak of "this"
solar system, or this Earth. We may, of course, speak
of "any one" as I did in the last paragraph.
Lee
This archive was generated by hypermail 2.1.5 : Thu Mar 06 2003 - 00:37:28 MST